StockSheild

Don't invest blind. Scan first

Comment

Inspiration

The idea for StockSheild was born from a real incident. A close family member received a WhatsApp message promising guaranteed 300% returns on a stock. Trusting the tip, they invested their hard earned savings — and lost everything within days. This is not a rare story. Every day millions of salaried Indians receive fake stock tips on WhatsApp, Telegram and YouTube, and many lose their life savings trusting them. SEBI investigated over 1,400 stock manipulation cases in 2023 alone, yet ordinary investors had no simple tool to protect themselves. We built StockSheild because every Indian deserves to know the truth before investing a single rupee — in 30 seconds, for free, in plain language anyone can understand.

What it does

StockSheild is India's first AI-powered stock manipulation detector built specifically for ordinary salaried investors. It works in two simple ways — a user can either enter any NSE or BSE stock symbol, or paste any suspicious stock tip received on WhatsApp, Telegram, or social media. Within seconds, our AI analyzes the input for manipulation patterns, fake SEBI claims, pump and dump signals, unrealistic return promises, and urgency tactics. It then generates a manipulation score from 0 to 100, lists every red flag detected, delivers a plain English verdict, and suggests exactly what action to take next. No financial jargon. No signup required. No expertise needed. Just paste, scan, and know the truth before your money is gone.

How we built it

StockSheild was built entirely using vibe coding as a solo project in under 24 hours. The frontend was designed and developed using Google Antigravity, which allowed us to rapidly build a professional dark-themed UI with live alert feeds, animated scanning effects, and a fully responsive layout without writing complex code manually. The AI manipulation detection engine is powered by OpenAI's GPT model, which was given a carefully crafted system prompt trained to identify specific manipulation patterns, fake SEBI credentials, pump and dump language, urgency tactics, and unrealistic return promises commonly found in Indian stock scam messages. Stock symbol data is fetched in real time from Yahoo Finance API using the NSE and BSE suffixes, giving users genuine market context alongside the AI analysis. The entire application runs as a single deployable web app with no backend database required, making it lightweight, fast, and accessible on any device including mobile phones — which is where most Indian investors receive suspicious tips in the first place.

Challenges we ran into

The biggest challenge we faced was getting the AI to consistently return structured JSON responses without breaking the application. In early tests, the model would sometimes return conversational text instead of clean JSON, causing the entire results panel to fail silently. We solved this by carefully engineering the system prompt to explicitly instruct the model to return only valid JSON with no markdown, no explanation, and no preamble. The second major challenge was the CORS error that blocked all API calls when running the app locally as a plain HTML file — something we had never encountered before as a beginner. After research we discovered the fix was simply running the app through a proper local server environment. The third challenge was Gemini API hitting its free quota limit repeatedly during testing, which forced us to evaluate and switch between multiple AI providers including Gemini, Groq, and finally OpenAI to find the most stable free solution. Each of these obstacles taught us something new about how real production applications are built — and made the final working version much more robust than where we started.

Accomplishments that we're proud of

Our biggest accomplishment is building a fully functional AI-powered application in under 24 hours as a complete beginner with almost no prior coding experience. When we started this hackathon we had never built a real web application, never called an API, and never deployed anything live on the internet. Today we have a working product with a professional UI, real AI integration, live stock data, and an actual use case that can protect millions of Indians from financial scams. We are proud that StockSheild is not just a demo or a mockup — it genuinely works. Paste a real suspicious stock tip and it gives you a real AI-powered verdict in seconds. We are also proud of how we handled every obstacle — CORS errors, API quota limits, JSON parsing failures — none of which we had ever encountered before. We debugged each one, found solutions, and kept building. But above everything else, what we are most proud of is building something that solves a real problem that affects real people. This is not a todo app or a weather widget. This is a tool that could genuinely protect someone's life savings. That feeling is worth more than any prize.

What we learned

This hackathon taught us more in 24 hours than we had learned in months before it. On the technical side we learned how to integrate a real AI API into a web application, how to engineer prompts that return consistent structured outputs, how to fetch live financial data from external APIs, how to debug CORS errors, and how to deploy a working application that anyone in the world can access from their phone. But the most important lessons were not technical at all. We learned that the best products are not built from clever ideas sitting in a room — they are built from real pain that real people experience every day. We learned that being a beginner is not a disadvantage — it forces you to ask simple questions that experts overlook, and those simple questions often lead to the most impactful solutions. We learned that vibe coding is not cheating — it is a superpower that lets anyone with a genuine idea bring it to life regardless of their technical background. And finally we learned that shipping something imperfect but real is infinitely more valuable than planning something perfect that never gets built. StockSheild is not perfect. But it works. And somewhere out there it might just stop someone from making the worst financial mistake of their life.

What's next for StockSheild

StockSheild today is a prototype — but the vision is much larger. The immediate next step is connecting to real time NSE and BSE data feeds so that manipulation detection is based on actual live market numbers rather than tip analysis alone. We plan to build an automated monitoring system that continuously scans all 5,000+ stocks listed on Indian exchanges and sends instant alerts when suspicious volume spikes, unusual options activity, or coordinated social media manipulation is detected — before the pump happens, not after. We want to launch a WhatsApp bot version of StockSheild so that any Indian farmer, teacher, or shopkeeper can simply forward a suspicious message to our number and receive an instant verdict in Hindi or English without ever opening a website. We are also exploring a browser extension that automatically scans any stock tip you see online in real time. On the business side we see strong potential for partnerships with SEBI, NSE, BSE, and financial literacy NGOs across India who are actively looking for tools to protect retail investors. Longer term we envision StockSheild becoming India's most trusted financial safety layer — the way an antivirus protects your computer, StockSheild protects your savings. Because in a country where 10 crore new investors entered the market in just 3 years, the need for financial protection has never been greater — and we are just getting started.

Built With

Share this project:

Updates